Auxiliary material for Quantification of factors impacting seawater and calcite d18O during Heinrich Stadials 1 and 4 Witold Bagniewski (1), Katrin J. Meissner (1), Laurie Menviel (1), and Catherine E. Brennan (2) (1) Climate Change Research Centre and ARC Centre of Excellence for Climate System Science, University of New South Wales, Sydney, NSW, Australia. (2) Department of Oceanography, Dalhousie University, Halifax, NS, Canada. Paleoceanography, 2015 Introduction Table 1 summarizes all the simulations performed in this study. In addition to the two simulations analyzed in the main text (fw20 and fwN, here named fw20-NPDW and fwN-NPDW), six more simulations were performed: two with different meltwater d18O ratios (fw30-NPDW and fw40-NPDW) for the same circulation mode, and four for a circulation mode without deep water formation in the North Pacific (fw20-noNPDW, fw30-noNPDW, fw40-noNPDW, fwN-noNPDW). Figures 1 ("fs01.eps") and 2 ("fs02.eps") show the changes in deep water formation rates and d18Ow anomalies for the two modes of circulation. In simulations without NPDW, AABW formation rate is reduced to about half of the rate obtained in simulations with NPDW, while NADW formation recovers about 100 years earlier. The differences in d18Ow anomalies for the two modes of circulation are generally lower than 0.1permil and may be assumed too low to significantly affect fit with paleoproxy data. Figure 2 ("fs02.eps") also shows the sensitivity of the simulated d18Ow values in the North Atlantic and North Pacific to the isotopic content of meltwater. Figures 3A and 3B ("fs03.eps") show the surface d18Ow and d18Oc anomalies due to the "circulation and climate change signal" (fwN simulation). The see-saw connection between the North Atlantic and the North Pacific is most visible in Figure 3A ("fs03.eps"), as d18Ow decreases in the surface North Atlantic by about 0.5-1permil and increases in the surface North Pacific by about 0.2-0.4permil. Figure 4 ("fs04.eps") shows the d18Ow and d18Oc anomalies in the Atlantic (A and B) and in the Pacific (C and D) due to the "circulation and climate change signal" (fwN simulation). Figures 5-8 ("fs05.eps", "fs06.eps", "fs07.eps", "fs08.eps") show the Dd18Oc anomalies from 36 sediment cores for HS1 and HS4, as well as Dd18O anomalies from the NGRIP time series, superimposed by simulated values of the same parameter at the same location (fw20 simulation). 1. Table 1: Summary of simulations. In all the experiments a freshwater (FW) flux (0.139 Sv) is added to the North Atlantic for 1800 years, after which a salt flux (equivalent to -0.2 Sv) is added for 1500 years to resume the AMOC. The d18O ratio of the freshwater flux added to the North Atlantic (NA) is shown in column 2. In some experiments a freshwater flux (0.15 Sv) is added to the North Pacific (NP) for 1800 years with a d18O ratio of 0permil. 2. fs01.eps (Figure 1): Time series of North Atlantic Deep Water formation (NADW, red), North Pacific Deep Water formation (NPDW, blue) and Antarctic Bottom Water formation (AABW, green) rates (Sv) in A) simulations with NPDW and B) simulations without NPDW (where an additional freshwater flux is applied to the North Pacific in order to stabilize the water column - see text for details). Bottom panels show a time series of the combined freshwater flux in the North Atlantic and North Pacific, where negative values represent an artificial addition of salt to force an AMOC recovery. 3. fs02.eps (Figure 2): Changes in average d18Ow in the North Atlantic (NA, red lines) and the North Pacific (NP, blue lines) between 30N and 60N for different d18O ratios of meltwater (fwN, fw20, fw30, fw40) at A) the surface for experiments with NPDW; B) 3200 meters depth for experiments with NPDW; C) the surface for experiments without NPDW; and D) 3200 meters depth for experiments without NPDW. Yellow area represents the time span of the freshwater fluxes. 4. fs03.eps (Figure 3): d18Ow anomalies (Delta, year 2000 minus year 0) simulated in experiment fwN, superimposed by paleoproxy anomalies for HS1 (squares) and HS4 (diamonds): (A) Model d18Ow and proxy d18Ow; (B) Model d18Oc and proxy d18Oc. 5. fs04.eps (Figure 4): Anomalies (Delta, year 2000 minus year 0) simulated in experiment fwN, superimposed by paleoproxy d18Oc anomalies for HS1 (squares) and HS4 (diamonds): (A) d18Ow in the Atlantic Ocean (zonal average); (B) d18Oc in the Atlantic Ocean (zonal average); (C) d18Ow in the Pacific Ocean (zonal average); (B) d18Oc in the Pacific Ocean (zonal average). 6. fs05.eps (Figure 5): Surface ocean Dd18Oc anomalies simulated in experiment fw20, superimposed by planktic foraminiferal Dd18Oc anomalies for HS4 (left) and HS1 (right). Paleoproxy records of HS4 have been shifted in time by the difference between the year corresponding to the end of HS4 defined in Table 1 of the main text, and 38.8 ka BP. Top panels show simulated anomalies of Dd18O in snow over Greenland at region between 28W - 18W, and 72N - 77N, superimposed by Dd18O anomalies from the NGRIP time series (NGRIP dating group, 2008). Yellow bars represent the time period for which the data was averaged to calculate anomalies shown in the main manuscript. Wherever the starting year for HS4 is different than 40.2 ka BP, the corresponding yellow bars are surrounded with a dashed line. Note that the model time scale is shown at the top. 7. fs06.eps (Figure 6): Same as Figure 5 ("fs05.eps") for cores in the Pacific Ocean. 8. fs07.eps (Figure 7): Same as Figure 5 ("fs05.eps") for benthic records. 9. fs08.eps (Figure 8): Same as Figure 5 ("fs05.eps") for benthic records.